THE ALBERT SHANKER INSTITUTE
WHAT DO WE REALLY KNOW ABOUT
HIGH SCHOOL DROPOUT RATES
& WHAT CAN BE DONE TO IMPROVE THEM?
A FORUM
WITH
ROBERT BALFANZ
RESEARCH SCIENTIST, CENTER FOR SOCIAL ORGANIZATION
OF SCHOOLS,
MARK DYNARSKI
SENIOR FELLOW, ASSOCIATE DIRECTOR OF RESEARCH,
AND DIRECTOR OF THE EDICUATION AREA AT MATHEMATICA
POLICY RESEARCH; PRINCIPAL INVESTIGATOR FOR DROPOUT PREVENTION, WHAT WORKS
CLEARINGHOUSE
MODERATOR:
RANDI WEINTGARTEN, PRESIDENT,
UNITED FEDERATION OF TEACHERS;
MEMBER, ALBERT SHANKER INSTITUTE BOARD OF DIRECTORS
THURSDAY, MAY 3, 2007
12:30 P.M. – 2:30 P.M.
NATIONAL PRESS CLUB
Transcript by:
Federal New Service
EUGENIA KEMBLE: Welcome. I’m Eugenia Kemble, the executive director of the Albert Shanker Institute. I just want to say a word about our forum series and what we’re trying to do. Al Shanker, who our institute is named after, was known to many as someone who was genuinely interested in good research — particularly in the field of education, where there is a tendency to find the research that supports your point-of-view and then marshal it as the reason you are doing what you already wanted to do. That wasn’t true of Al.
Now, of course, Al had certain things he believed and things he wanted to accomplish. He believed that good teachers were good for kids and for good schools, and he wanted to help them do their jobs better. That was the bottom line. But beyond that, he was constantly searching for ways to come up with the best policies for education — for public education in particular.
I just want to mention one example where attention to research led Al Shanker to surprise everyone. It also demonstrated how serious he was about an honest appraisal of the facts. This was in 1983 when A Nation at Risk came out, a report that was pretty critical of public education. And everybody anticipated that Al, like many others in the education world, would simply attack the report. He didn’t do that, because he felt that there was much in the report that was worth looking at and learning from in the AFT’s quest to improve schooling. And so, we’ve tried to pick up on that legacy and create a forum series, where we would discuss the best research we could find on key education topics — especially those where there is great variance between what the research indicates and what public policy really is, or where a pressing issue has a research base that is woefully thin.
As always, the researchers we will
hear from today were selected through a kind of informal peer review process,
where we ask top people in a given field who they think the best researchers
are, then ask those people, and so on. Our moderator today is Randi Weingarten,
president of the United Federation of Teachers, the AFT’s New York City affiliate, which includes not
only teachers, but teacher aides, and retirees, and which is — I don’t think
that this has changed — the largest local union in the world bargaining with a
single employer. She is also on the Shanker Institute board of directors and is
a key leader of the
We’re really glad Randi could come
today all the way from
Also, before I turn this over, I just want to also tell you that Antonia Cortese, executive vice president of the AFT, is here. She is also on the Shanker Institute's board of directors and is really the education policy czar at the AFT. And I also want to thank Burnie Bond who put this whole forum together.
RANDI WEINGARTEN: My first task is to make sure the microphone and the salad don’t meet. Thank you, Genie, for the introduction. All it says to me when I get introduced is that I can’t actually sit still, and go from place to place. (Chuckles….)
I wanted to do this forum for a
couple of reasons — first, because of how important the Shanker Institute is to
the attempt to have reflective, interesting, important conversation about topics
in education. All too often in education these days, everything comes down to
sound bites and vituperative comment in between the sound bites. And so what
the Shanker Institute really does is it tries to move away from the sound bites
and actually tries to figure out how we improve the education of all of our
kids in urban
On this issue, as Genie said, this is a very – as I have started reflecting about what public education can really do K through 12, the whole notion of dropouts and when you identify dropouts and what happens in terms of dropouts, we’ve been thinking about a lot in New York. And in fact, one of the reasons that we are so focused on middle schools is because we see that as a lynchpin, frankly, for how you solve this problem. But first, you have to figure out the dimensions of the problem, and that’s part of what we’re doing today.
Everyone agrees that even one
dropout is one too many. But there are
more disagreements than not over just how many dropouts there are. Getting a better read on which kids and how
many are dropping out and why is pretty critical because you can’t really
design an effective solution, a systemic solution, if you don’t have a good
grasp of both the scope and the particulars.
And as we know, those estimates are all over the lot, and some people
use them for political reasons. We’ve had
estimates as low as a graduation rate – I’ll get to this in a second – of 30
percent in
But, the one thing I had to learn
is that the dropout rate is not the flipside of the graduation rate. Even I, for awhile, was making that
mistake. Not every one who, quote, “doesn’t
graduate in four years,” end quote, drops out.
In fact, many youngsters – I’ll take
So, what you look at is you also
look to see, not just what the graduation rate is four years, but, at least in
Dropouts come in very many varieties. The programs to catch these kids up also vary hugely and I think part of our dilemma is we know very little about what is effective. We know, for example, that it takes more than smoke and mirrors to get kids to graduate and to reduce the dropout rate. We know that it takes time, money and skill but what are the systemic programs that we can use to try to help?
So, for example, the dropout rates – what we know is dropouts rates are highest among those populations of students who are most in need of a good education. That is: poor, immigrant and minority students. Second, as I said before, it doesn’t appear – actually, I did say this yet. It doesn’t appear to be much evidence that the adoption of tougher standards and exit exams have driven up the dropout rate. Even in the worst case scenario, the opinion that the graduate – there is a consensus of opinion that the graduation rate has remained pretty steady since the early 1990’s when standards-based reforms began to take hold.
And third – and this was what I was trying to say before – we all agree that few high schools, big or small, are equipped to handle the very real needs of students who are at risk of dropping out. They don’t have intensive academic interventions for kids who are struggling academically. They can’t draw on the support of the social service system to help with students whose family or community environments are part of the problem. They don’t have the capacity to offer career counseling that could help students understand the tangible benefits of staying in schools.
So essentially, we don’t really have – and that’s what one of our speakers is going to talk about as well – we don’t really have a consensus of opinion right now on what, you know, what are the five things that would really work to help ameliorate the dropout problem and to help create a higher and higher graduation rate in terms of the four years.
So, that’s why we need this panel and why have two fantastic researchers because, first, we’re going to hear from Bob Balfanz, who will lay out how using different data sets and asking different questions about data has resulted in different numbers being throw out – press, policy circles – with very little understanding of what they really mean. And then we’re going to hear from Mark Dynarski who will provide an overview of the research on the policies and the practices that are likely to help the most. And before I go to Bob – and then we’re going to open it up for questions. Now, what I’m going to ask after Bob speaks, and also after Mark speaks, is if either one of them want to just respond directly to what each other has said and then I’m going to open it up, as I said, after both of them speak, to questions from the lunch attendees. But remember what you have to do: identify yourself.
Let me just say one thing about Bob
and one thing about Mark. Bob is a research
scientist at the Center for Social Organizations of Schools at
Mark: Mark Dynarski is a senior fellow, Associate Director of Research and Director of the Education Area at Mathmatica Policy Research, Inc. where he has worked since 1988. And prior to that, he was professor of economics at UC-Davis. Dr. Dynarski – both of them are doctors, obviously – have a national reputation for his work in education policy, particularly in evaluating programs for at risk children and youth in school community partnerships. And he, too, has published numerous articles on these topics.
So without further ado, can we start with Bob and then Mark? Thank you very much, both of you, for doing this.
BOB BALFANZ: Thank you and good afternoon. I want to thank the organizers for the faith they have in us. Here’s what they said, in ten, fifteen minutes, maybe, give us a dispassionate analysis of the graduation rate crisis, hit the hot button issues and bring us to reauthorization of NCLB. So, instead of talking really fast, I’m going to be selective and leave some room for the questions.
And what I’m going to do is try to give just the very brief history of graduation rates since 2000 to try to answer the question of why is this so hard? Why are we here talking about in 2007 how to measure graduation rates? Isn’t that the fundamental measure of schooling? Isn’t that the fundamental outcome? Do you graduate or not? Why would we be here in 2007 saying, well we’re not sure what it is?
Well here’s the story. I bring you back to the halcyon days of 2000, right? Goals 2000 was to be achieved then. In Goals 2000, the goal was 90 percent graduation rate for the nation. And if you read the press around 2000, you might think we’re pretty close. The U.S. Census Bureau would come out reports saying graduation rates are up, they’re almost at 90 percent. Minority graduation rates are up. If you looked at district data, you would see them report dropout rates of two percent, three percent, four percent. You would say, that seems understandable that maybe two, three or four percent of kids won’t graduate from high school today. I mean, kids do have to work. Some kids do get alienated. That doesn’t seem like a crisis. So, part of the reason why is that people didn’t really think it was a problem. So they didn’t think it was a problem; they weren’t really trying to think deeply about how to measure it.
But some discordant notes were
beginning to be hit. There were some
rabble rousers down in
And this was about the time we were
working in a lot of high schools and we started noticing it in a very different
way. We would be in high poverty, low
performing high schools and they looked the same to us. It was
So we started asking ourselves how many of these high schools might there be. This is how I sort of got in through the backdoor. And first we looked in the 35 biggest cities and we found about 200 of them. And what we used as our marker was just an indicator. We said, let’s compare 9th grade enrollments to 12th grade enrollments four years later.
Now we know this is not a graduation rate. We know this is just, at best, an indicator but what it’s pretty good at, we think, is distinguishing a school with really low graduation rates from one from high graduation rates because it would seem to reason that if you have high graduation rates, then your freshman class out to be about the same size as your senior class, unless there’s high migration, which you could know about because that would be a knowable – a big plant closed, a big plant opened; I mean, things that happen to cause big migration are knowable. But absent that, it should be in balance. And if you find a freshman class that’s routinely two or even three times as big as the senior class, well, that’s probably a school that’s got a problem. At a minimum, it means the kids are fleeing the school. Maybe they’re going to other schools in the district and graduating. So at a minimum, it’s not holding students.
And, the reason we had to use 12th grade and not graduates is because, again, believe it or not, the federal government or the state governments at the school level don’t collect the number of graduates. In the federal data collection system, that’s only collected at the district level and it’s lagged a year. So you have to wait another year to get it. So you could only at the district level, could you even compare 9th grade to graduates. So we had a data problem too. I mean, we couldn’t get a uniform way at the school level to even estimate the graduation rate.
About the same time, a bunch of
other folks did take it to the district level, like Chris Swanson at the Urban
Institute and Jay Greene at the Manhattan Institute and Rob Warren,
But what was true though is – and they all try slightly different ways to deal with that; some used 8th grade data, some averaged 8th, 9th and 10th grade data, some just looked over two years so you wouldn’t have a migration problem. But it is true however they are all finding lower rates than the nationally reported rates.
And that, again – let’s take one step back and look at the census and why maybe that might not be perfect. The census is based on asking people, calling them up and saying, did people in your household graduate, essentially. And, so first there’s the sampling problem. I mean, you know, in high poverty neighborhoods we know there’s a census sampling problem to begin with. And second of all, people – when you say graduate, they mean different things. Often they equate a GED with graduating. Often they equate being at high school for four years for graduating because people have found that even if you ask people to self-report what grade you’re in, they often get it wrong because they’re saying how many years have I been in high school. I’ve been in high school for three years, I’m an 11th grader. Credit-wise, they may in fact be a 9th grader. So, there is reasons why self-reporting in the census could lead to an overestimate especially just in certain areas.
So that was the background on the measurements. Then NCLB came along in all of this, right; the next part of the story. And at that point, there was a great desire to have balance to test scores because there was a fear that if we just have test scores as the accountability measure, the obvious thing at the high school level is to push the kids out that might have low test scores. Now the extant that schools that really do this, I think, is overstated but I think it does happen in certain places. It’s almost at the school level that sometimes a over exuberant assistant principal knows how to work the system, but it is a legitimate concern and they wanted to have grad rates balance that. So they got grade rates in NCLB but because there wasn’t a good way to measure them yet, they didn’t sweat the details. And there’s a lot of politics behind this, too, which we’ll skip over and we can talk about in the discussion.
So they sort of left all the details up to the states. They said you must have grad rates but we’ll let the states figure out what your goal will be, how much students should improve and how you’re going to measure it. And again, the states – not in a nefarious way, I don’t think – but had different systems in place, had different ways of looking at it, had different capacities to invest in a whole new system.
And so we got a real range. We got some states saying, well, any
improvement counts, so if you go from 50 percent to 50.5 percent, you’re
okay. And I remember that Dan Losen of the Harvard civil rights institute called
Again, because we didn’t have a
good measure, we got lots of states going in different directions. So even today, we can’t look at graduation
rates in NCLB across states and have any idea if
This led to how can we do this better. And two more groups come into play here. The National Governors Association gets into the fray and they work out what’s called the NGA compact. And in that, they try to get an agreed upon graduation rate that, in principle, all the governors signed onto. And also at the same time, the U.S. Department of Education, in a parallel effort, came to more or less the same rate. And this is a pretty logical one. I think this is where the direction that we’re heading in the future. It’s just going to take a while to get there. Which is, simply put, you do want to start with your first time freshmen – because that’s you sort of beginning class; that’s the who comes into high school and now it’s your charge to graduate them – and you do want to measure how many of them graduate within four years, because on time does matter. Now you do want to also measure how many of them ultimately graduate also. So we can talk about five year or six year rates. That needs to be there.
And the two things you need to keep
track of to make this fairly accurate is transfers in
and transfers out. Now, transfers out is
the tricky part because this has been the black hole of dropout and graduation
rate accounting because the bane of accurate measure, and also where you could
play games if you want to, is something called the non-drop withdrawal, which
means, you left the school, but you didn’t drop out and this open to definition
sometimes by the school, sometimes by the district, sometimes by the
state. So for example, in
In other states, cities there’s the catchall for we don’t know, which is whereabouts unknown, is an example. Because many kids who dropout don’t come in and say, I’m dropping out tomorrow and here’s why, they just stop showing up. And this, actually, puts a real burden on schools and it’s another thing to keep in mind about this is that it’s ultimately the school that’s got to ascertain why that kid isn’t showing up anymore.
Now, in some schools that are big and have lots of dropouts, they actually have a function for this. There’s someone called the drop secretary who, sort of, her job or his job is to try to call the kid or the family and on a checklist check why they dropped out.
So remember when you ultimately see
all your graduation and dropout statistics, think of it going all the way back
to the busy person in the school with a telephone and a checklist and that
should always give you a slight pause in ultimately how accurate, you know,
some of these detailed explanations of why we drop are. I mean, they try hard, but, obviously, phone
numbers are hard and what do you do when someone picks up the phone, I’m his
cousin and he moved to
And so at the actual point of action, this becomes quite difficult, but that’s why there’s this big loophole with non-drop withdrawals because you can make it so big that no one officially drops out. And that’s when you often see these incredibly low dropout rates and these incredibly high low graduation rates at the same time. And you say, well, if they’re not graduation and they’re not dropping out, where are they going? They’re withdrawing. And so that’s another key thing to keep in mind and people have been pushing is that even if we ultimately get a common measure for graduation rates, say the NGA compact, if there’s not agreement on withdrawal codes, they’re still not going to be comparable. And this gets, you know, very technical, right? It’s like, oh my god, this is so bizarre and hard to follow, but if those details aren’t right, the ultimate measure isn’t going to be right.
Two more things and then we’ll let Mark get in here. You’ve heard a lot, maybe, individual identifiers, which is a thing where everyone keeps what we ultimately need is that we need to have a clear code that lets us follow students through schools, across districts and ultimately across states. There’s a whole national data clearinghouse – they have a good title; I’m not using it – that’s pushing for this. And ultimately, that seems like a very useful thing because we can really do good fundamental research in tracking, but what people will tell you is, well, that’ll 2015 to get there and we can’t do much about graduation rate until then. And that seems a little too long to wait, right, because we’re in 2007 and back in 2000, we couldn’t measure it. And so that’s another thing to keep in mind is, yes, individual identifiers are important but that, in a way, becomes its own loophole. Until every state has them – and then they need five years of data until they can measure their rates – so that gets us back into this messy area of estimates and people don’t want to get there, but I think for the interim we do have to consider using some sort of estimate – not as an absolute measure; not saying this the graduation rate – but as a way of follow growth, right, because that’s easier. Even if the rate’s off by five or 10 points, the growth should be easier. Leaving some room for the great confidence interval of AYP, is useful here. (Laughter.) Then keep that in mind.
And then the last thing, which
segues into Mark, is this whole sort of growing area of predictors – and Randi
was talking about this – which is that the other thing we found in our research
– and this will be my final point to tie things in on why the graduation rate
is difficult to measure – is that our basic findings are is that the graduation
rate crisis, the dropout crisis, is not evenly distributed across the
country. It’s not the same in
But what this means is if you put a graph of, sort of, graduation rates by schools, it’s likely that those with 90 percent would have the biggest number, okay, because in many communities, most kids graduate. And so, that’s one reason why this crisis isn’t felt everywhere and it’s only down here that you would have maybe 15 to 20 percent where many kids don’t graduate. So that’s one reason why it’s hard to measure. It’s not equally distributed. But because of that, it’s largely driven by school failure; the students failing in school, not attending, not earning credits, not being promoted. That actually makes it quite predictable because in the mass, there’s going to be a group of kids that will always drop out for life events: work, pregnancy, alienation. They’re very hard to predict, but in these schools where lots of kids drop out, that’s not the majority of dropouts.
The majority of dropouts are kids failing in school and failing in school follows a very predictable path, which means you can have early identification in the middle grades because that’s when we actually is when kids – this is transition to adolescence – and they’re sort of making an independent decision to be engaged in school, right? In elementary school, you’re told, follow the leader and if you’re good at that, you succeed. But when you’re an adolescent, you start making your own decision, right, and so they’re sort of re-upping their commitment to school and if that doesn’t happen, that’s when they sort of fall off the graduation track.
Now, final thing is that kids are remarkably persistent. They don’t just, you know, wait until they’re 16 and walk out the door en masse. They actually keep trying, they keep repeating, they keep failing. I mean, it’s a give and take. In other words, there’s probably two sides of their head – yes I can succeed; no I can’t – talking everyday and ultimately they get worn down. But again, this means there’s time to intervene. So, we can both know where many of the dropouts come from, we can find them early, there’s time to turn it around and, hopefully, with this really good setup, Mark will now tell us that there’s something we can do about it. (Laughter.)
MS. WEINGARTEN: Thank you, very much.
MR. : But what if he can’t? (Laughter.)
MR. BALFANZ: Get someone else. (Laughter.)
MARK DYNARSKI: Thanks very much. I want to thank Bernie and Eugenia and Randi and the institute for inviting me. There is a handout – a set of handouts – in the packet and I realize I forgot to number them but they are all pretty distinct, so there should be no problem in finding them. And so, as I talk there will be various visual exhibits that people can refer to. But what I want to do is talk some about he national scope of this dropout problem by looking at NCES data which I don’t think have some of the kinds of issues with the dropouts rates that have arisen with local school districts. Because these are from national data sets, the levels might be off, but the trends are probably still pretty accurate.
We want to look at some results of having reviewed, having done an extensive review of the literature about the effective of dropout prevention programs, which is part of the works for the What Works Clearinghouse, which is funded by the Institute of Education Sciences. I’ve been working with them for a couple of years. I think in October, we began releasing reports onto the What Works Clearinghouse website about the effectiveness of various interventions and those are going to be summarized here. And that work continues. And then I can offer some observations about where I think we are at the current time with this program development side of thinking about things and some suggestions on where we might go next.
So with that, let me first just mention the dropout problem. It’s incredibly stable, actually. Since about 1972, it’s slid very gradually – and the third slide here shows you this and I think it’s also in the red packet from an NCES report. You can kind of detect the downward motion of this dropout rate. So, these are essentially kids in the 16 to 24 year old age range who we say they are not currently attending school and they’ve not completed high school. So, it does slide down very gradually but I drew a line here on the slide at 1990 and it’s almost imperceptibly lower from 1990 to the current day. So, over the course of 17 years, it has barely moved.
It is much higher for Hispanics than it is for the total and for whites. And in fact, that Hispanic number, because it includes both immigrant Hispanics and native born, is somewhat of a – yet to take into account the fact that immigrant Hispanics will have had a much different educational background. But if you look at native born Hispanics, the dropout rate will be about 15 percent. So even native born, have three times the dropout rate of, well, white students.
Now, if all the dropouts just sort of where stopping out for a while, taking a vacation from high school but returning, you would end up in the situation which Bob alluded to. You can get these complex dynamics going on the in the comparing the dropout rate to the completion rate because if they all take a year off and come back to school and ultimately complete, you can have a 100 percent completion rate while you still have a segment of your population that could dropout at any point in time.
The next slide shows, however, that that’s not what’s happening. Many dropouts are in fact stay outs and, in fact, a recent brief from NCES, they looked at the data from NELS for kids who were now eight years past the point when they would have normally graduated from high school if they had been on time, so they were up there around 25 or 26 years old and they were looking at their high school completion. So this is about as late as adults normally would complete high school and about a third of those kids had still – basically, if a student ever drops out while they’re in high school, two thirds of them will complete by the time they’re 26, which interesting. Not many of them will get high school diplomas. Mostly they’ll get alternative credentials like a GED. But one third will just be permanent dropouts. And that’s obviously the kind of student that we’re most concerned about because, frankly, we just all worry that these folks have just stepped off the economic ladder to well being. They’ve just completely done themselves in, so to speak, with respect to any kind of vital future.
This is all happening in the context which the next slide shows. This is K-12 education spending from the NCES since between 1990 and 2002 in real per student dollars, so this is actually how much is being spent in constant dollar terms. And since 1995, this has really taken off. Now, I admit in looking at the slide, you can’t tell whether they’re actually spending this money on high school or middle school. If they had spent all the increase on elementary school students, it wouldn’t do much good if you were already in middle school. I have a feeling it’s more general than that, but the main point I was trying make is that the dropout rate is not moving at all in a context when we’re pouring a lot more resources into education. And so, there’s obviously – this creates consternation that why is it no working?
In fact, if you go back two slides, I want to just mention something very briefly here. This dropout rate upticked in the very last year of the data collection in 2004 and we do not know whether this is a statistical anomaly or something which might be occurring as a result of higher standards and minimum competency tests in high schools. We need another couple of years of this, but that rate to uptick in all race, ethnicity classes is a little bit worrisome.
In 2002, when the What Works Clearinghouse was created by the institute, dropout prevention was on its list of the initial five or six topics that it began focusing energy on. I began working with the clearinghouse in 2005, so we’ve now had a couple of years of experience. Essentially what the clearinghouse is trying to do, just to set notions here, is it’s not doing primary research primary research on its own. It’s gathering up all that it can identify out there about the topic area and, in particular, it focuses on studies which show causal evidence of effectiveness. Now by causal evidence, I essentially mean that this intervention yields some kind of outcomes when you put it into practice. And so, a study which has a control group or a comparison group will likely be a causal one. It won’t necessarily be a strong one, but essentially what the researcher is probably trying to show is that the intervention generated some kinds of outcomes.
There’s lots of other research about dropping out, for example, which might just take the NELS and look at various kinds of relationships or correlations that are at work. And in fact, there’s been some well known articles, for example, by Valerie Lee who show that there’s a relationship between dropout rates and the size of a high school. But those kinds of studies fall outside the purview of the clearinghouse because the researcher is not trying to show the effectiveness of an intervention, they’re trying to demonstrate that a basic relationship exists. And so, keep that in mind as we move forward.
So we had to set some parameters
for the review. And so we took as our
presumed audience that a state taskforce was charged with the task of reducing
dropout rates. And the reason we chose
that is because dropouts are, in fact, targeted by job training programs, by
local non-profit organizations, by state agencies, as well as by schools. Mostly by schools, but we recognize that
there was research about all different approaches to try to reduce the dropout
rate. Imagine, if you will, the
So we chose this very broad kind of approach to identifying the literature so that we weren’t focusing solely on K-12 approaches. And the interventions had to actually have as one of their primary purposes to keep kids in school. And that sounds sort of painfully obvious if you’re looking at dropout prevention, but in fact there is a much larger research about kids in high school which focuses on broad risk reduction strategies like teen pregnancy and drug use and anti-gang programs, anti-violence, which might cause the dropout rate to decline, but the programs themselves are not actually being structured so as to reduce dropping out, per se. They have a different focus. And so, to avoid being overwhelmed, we faced a dilemma here between looking at a relatively small literature versus an absolutely staggeringly large one, we opted for the smaller one.
And then we also had to decide what kinds of things did we want these interventions to do because there’s all different kinds of outcomes that an intervention can try to focus on. And so we identified three different domains. One: are the interventions keeping kids in school? Are they progressing kids through school? And then, are they getting them to complete school? So there’s a logical interrelationship of the three domains. You have to stay in school to progress. You have to progress to complete. Completing will then, logically, be the hardest one, the one least likely to observed simply because you have to go through the other two to get there.
We then began the literature search. We do, you know, the standard kinds of e-searching and Google Scholar and the like. We find 4,000 hits on various kinds of terms related to “dropout.” This is not necessarily good news because 4,000 articles is a lot of literature to try to read. But then, some of the articles will prove to be things like somebody studying dropping out of yoga programs. We realize, okay, good, we don’t have to worry about that. And then, dropping out of college, which is an entirely different issue, but one that we could sidestep.
Then we noticed that a lot of the literature is, in fact, policy briefings. The transcript of today’s talk, for example, would fall into this kind of hit; testimony to various congressional committees and the like. But there wouldn’t necessarily be a real study underlying this, it would just be talking about dropout prevention. And so we could screen all that out. So we continued these kind of screening exercises and the next thing you know, we’re down to 85 articles which are actually articles which purport to look at the effectiveness of an intervention to reduce dropping out. This is not a big literature anymore. It sounds large when you start, but when you actually focus on programs that reduce dropping out, it sharply drops off.
Then, you do the last cut. You impose the What Works Clearinghouse standards on it. Now the standards are actually separate from the review areas. The standards were created as a core which applied to all review areas, but they essentially give weight to experiments and to quasi-experiments. Quasi-experiments are essentially they don’t use actual random assignment, but they create a comparison group of kids through some kind of algorithm that we can be confident in. And now it’s 15 to 20 articles. We’re not done reviewing all this but now it’s not very much literature at all.
Now, it is definitely the case the standards matter here a lot. If you relax the standards, you’ll admit other kinds of research in, but at the same time you should have less confidence in what the other research actually is telling you because these standards are essentially designed to screen out potentially misleading findings, the ones in which they’re saying the intervention is effective but, in fact, it’s entirely an artifact of the design. It’s also very common we found for studies to be multiply cited seven different ways and it’s actually the same study when you go back to it.
So there’s a page here called findings which is essentially just a screen shot from the What Works Clearinghouse webpage. And so, I don’t know how visible it is here, but essentially, we’ve identified these eight programs to this point for which we are very comfortable saying that the evidence passes standards. Now, that doesn’t mean the evidence shows that it’s effective. It means that basically it just merits inclusion in the clearinghouse because it’s well done research. And there are also other kinds of aspects of it such as the researchers were able to maintain their sample sizes and the like. But, in looking at that list, the code is essentially: a plus with a question mark next to is potentially positive.
There’s one intervention here which is positive. Now positive is basically unassailably positive. There are two different replications of an intervention called Check & Connect. Both of them were experiments, so they did two rigorous studies that were separate and they showed the same findings. So that, in the clearinghouse, gets it the highest ranking. If you only did one experiment and it showed a positive, it would only count as potentially positive because the clearinghouse standard demands a replication in order to give it the highest vote of confidence.
But partly at issue here – and this is something I’m throwing out there for discussion – this is a very underresearched area. For 20 years, you’ve had a social problem which has not moved one whit and it’s pretty big. And if the legislation that was just introduced, if those numbers have any veracity in them – we’re talking about hundreds of billions of dollars of lost earnings, of additional payments in welfare and the like – and basically, the clearinghouse review has identified eight things which might work. And our window is a 20 year window, so, I mean, our efforts are continuing but I don’t think it’s going to go much beyond a dozen here.
So, just let me put out there that
you’re going to be picking through some relatively modest sets of
findings. I think they’re good
interventions, but you don’t have a lot of stuff here and in particular, you
don’t have replications. Scientists,
researchers, we love replications. The
point of replication reduces risk. If
you’ve only done an intervention once, in one school district, with one set of
kids and you have positive findings, what happened in
So, I tried to – if you jump back a couple of pages – to identify which program features actually worked. Well, this turned out to be a little bit frustrating, so I broke it. Middle school and high school; do you have an adult advocate who’s working with the kids? Do you have a small school? Do you have thematic or accelerated curricula? Do you have various kinds of elements which have been distinguished as potentially contributing to effective dropout prevention programs.
Now, there is a danger in this exercise which is that the ineffective programs often had poor research about them, so don’t even seen them here. So, I’m only looking at the effective programs and I’m looking at their features but I’m not seeing any kind of preponderance here that said that the effective programs all share a feature. In fact, they’re all over the map. So, there’s nothing that clearing emerges here as a panacea element that programs must include.
In fact, I should just mention,
there’s one called financial incentives which seems – I should probably explain
that. It sounds kind of – it’s not
really an intervention. These were
essentially – there was a program operated by the welfare agencies in
Versus something like Check & Connect, which is very much – it’s an adult advocate working inside a school. Check & Connect is one of these programs which named itself perfectly. They check with the kids. The connect with social services and with people in the school. So there’s an adult advocate who acts as the swing person, makes sure the kid’s progressing. If they’re not progressing, they do something about it. It originated with the special education community but the notions of how to operate a program like this are very broad. And in fact, the ALAS, which is Spanish for – I forget what – but it operated under the same principles in a school that was heavily Latino in L.A., but it’s just like Check & Connect only it was operating for Hispanic kids and so they called it something different, but it shared the same principles. They both had pretty large effects in terms of just their ability to move outcomes here.
Let me end on these considerations. Again, not much evidence; disappointing. We wish there were more. Now, people have accused researchers of always wanting more research. Unfortunately, that’s exactly where I find myself here. I mean, I used to say there’s so much good research done here that we just don’t need to do anymore and we could just move onto the operational lessons but, you know, if you’re really talking about eight or 10 programs for a country of this size and with this size social problem, this is not very many things. It’s a starting point, however. It’s a glass half full kind of thing. These are better – these are tested interventions and they’ve actually been found, most of them, to be effective, so that’s the place I would go.
Two considerations: why is it that so many kids regret dropping out and do it any ways? So, just in terms of – we’re trying to go back now to program design. So I was hypothesizing, well, they think they’re going to do better than the other kids around them that they might have heard about who dropped out and then didn’t fare so well. So there’s the whole notion that you’re invulnerable which is very consistent with being 16 and 17, okay. What then happens is over the course of the next year or two they realize that the job they had isn’t going anywhere. They’re not getting benefits ever. They want to have – they get older, they want to start a family and the like. It ain’t going to work out. So then they realize they need to go back to school. But for some reason, they convince themselves that it was okay to dropout.
One of the findings of in Civic Enterprise’s “The Silent Epidemic” is that nearly universally, dropouts said, well, that was a mistake. So I’m just thinking, is it something about these programs where they somehow get to these kids and they’re able to convince them that it’s a mistake before they do it. So, that’s something useful to think about.
And then the notion that Bob talked about: can you predict dropouts and focus services better? Well, there are all kinds of public service messages and the like which aren’t really very expensive in reaching out to kids. So, if you’re to reach 100 million people by putting a public service message on ABC, you don’t really have to worry about this problem. Just throw it out there. Kids who aren’t going to drop out, they see it anyways. They ignore it. Maybe it has some influence on those who do.
If you’re going to spend a lot of money, though, you really want this targeting to be pretty tight and Check & Connect, for example, is something like $1,500 a kid year. So, in this case, if you try to focus on the hundred kids in your school who are going to receive these services, you really want it to be the hundred who are going to drop out. That maximizes your program’s ability to have effects. The problem is, it just proves to not be very easy to find the hundred kids. And if you make mistakes, that means some of the kids you’re serving wouldn’t have dropped out and you missed some kids who are going to drop out and they’re still out in the general population, so to speak. So, this inability to categorize kids very exactly really undermines program effectiveness.
The other thing is, thinking about high schools, this adult advocate role. I think it really amounts to a communication channel that’s opened up and schools often do not have these kinds of communications channels because the kids move – every 15 minutes, the move to a new adult. And they have a guidance counselor – the guidance counselor has got a caseload of 400 – is often devoted entirely to things like getting some kids to college and writing letters and the like.
But having the an adult advocate in the school provides the teacher with somebody to talk to to indicate that so and so is having trouble. And this is very good information, I think, in general which is probably underexploited in schools, which is there’s attitudinal changes. There’s – teachers can see these things. An experienced teacher just knows when a kid’s had a bad weekend. Teachers actually can tell when a kid was the dad that weekend rather than the mom. They just know these things. And so, I think the ability to exploit what teachers know in a kind of surveillance way – I don’t mean to use the work negatively – but just the idea that the adults know what’s going on with the kids to a degree and if they could pass sit back to somebody who could do something about it, I think that’s a place to start. And I think that Check & Connect, ALAS and other uses kind of adult advocate approach are trying to use that kind of information and communication in order to resolve some of these issues.
I mean the most disappointing thing I heard when I was – this was a focus group some years back – a kid dropped out – he said he dropped out – because a teacher pushed him in the hallway. And so, you know, you’re in a focus group setting so I said, well, why did they do that? I don’t know, I don’t think he liked me, though, he said. And it just seemed like the strangest reason to have dropped out. Couldn’t somebody have stepped in and simply said to the kid, okay, why don’t we just talk to the teacher and we’ll see if – the teacher might have – it might have been an accident or there might have been some prior incident.
But high schools don’t really have intervention like this where people come in to just mediate things. And these are still young people, so they do things like they overread situations. And so the kid drops out of school because somebody bumped him in a hallway. I mean it was a really distressing kind of notion that it was actually triggered – such a serious life event could be triggered – by such a minor thing, but because they don’t know how to handle situations, and there’s no one there to help them handle it, it led too far.
So, let me stop there. I hope I’ve thrown some formative thoughts out there and then, discuss from here.
MS. WEINGARTEN: Thank you. Let me, as Bob is collecting his thoughts, let me just go around the room to see if people would just very briefly introduce themselves.
BURNIE BOND: I’m Burnie Bond from the Albert Shanker Institute.
EDWEARD MUIR: I’m Ed Muir from the AFT research department.
RANDALL GARTON: Randy Garton, formerly the Shanker Institute.
GERALD SHROUFE: I am not from the Shanker Institute. (Laughter.) Gerry Shroufe from the American Educational Research Association.
FRANK MURRAY: Frank Murray from the Teacher Education Accreditation Council.
VICKY THOMAS: Vicky Thomas (ph) from the Shanker Institute.
MR. : (Unintelligible.)
SYLVIA SEIDEL: Sylvia Seidel, National Education Association. I have the desk -- (unintelligible) –
CHRISTINA SAMUELS: Christina Samuels, Education Week.
RICHARD VERDUGO: Richard Verdugo, NEA.
THOMAS DIAL: Tom Dial, NEA.
JAMES MOORE: James Moore from the District of Columbia Public Schools.
JULIAN
ALISSA PELTZMAN: Alissa Peltzman, Achieve.
SHARI BROWN: Shari Brown,
KATE BANNAN: Kate Bannan, the
Knowledge
HOWARD NELSON: Howard Nelson, the AFT.
MIRIAM ROLLIN: Miriam Rollin, Fight Crime: Invest in Kids.
GISELLE LUNDY-PONCE: Giselle Lundy-Ponce, AFT.
MS. : Doc Hurzmark (ph), AFT.
SUSAN LEE: Public Education Network.
KYOKO SOGA: Kyoko Soga, Council of the Great City Schools.
JULIA LEDA: Julia Leda (ph), Council of the Great City Schools.
AMANDA HORWITZ: Amanda Horwitz, also Council of the Great City Schools.
DANIEL PRINCIOTTA: Dan Princiotta, National Governors Association.
TONI CORTESE: Toni Cortese, AFT.
JOHN EVON: John Evon (ph), Economic Policy Institute.
MS. WEINGARTEN: So, Bob, do you want to start us off or should we open the floor?
MR. BALFANZ: Sure, I’ll add one question and comment back to Mark to see if we can get agreement or disagreement on this. What I’m going to propose is there are actually two dropout problems in our nation, one which is pretty tough to diagnose and pretty tough to intervene, and the other which is pretty easy to diagnose, and difficult but possible to intervene in.
So I think there is two broad classes of dropouts. One is what I’m going to call, for lack of a better word, life events or inside the student’s head. Kids that drop out of school because something happens in their life, or sort of idiosyncratic decisions they make about interactions with school. So it’s you’re a kid who got pushed, is an example of it. It’s someone who has to feel they have to work. It’s someone who doesn’t feel safe going to school. It is someone who is just bored and frustrated and thinks a GED is an easier way to do it. It is someone who might be very – we had a student in the squad – (inaudible). He thinks he is a spoken-word artist; he is very gifted. He thinks he can go out and make a living for himself. And even though he has high skills and high grades, he just walks out.
Those are tough to predict because it’s about stuff happening inside someone’s head, right. And that is really hard to understand and hard to identify. And some kids that get pushed are very resilient and won’t think about it until the next day. And other kids, it changes their lives.
But there is this other group of kids that I think drop out, and I think in some schools, it’s the dominant number of kids, especially schools where lots of kids drop out, and they are dropping out because of school failure. They are dropping out because they have had poor prior preparation and they haven’t become engaged with schooling. And the schools they go to are pretty much overwhelmed. They serve a high concentration of these high-needs kids and are basically under-resourced to do the job. So they do the best they can, but they engage in triage.
And so what happens is, for these kids, dropping out is incredibly predictable. Their attendance falls off. They might – and they might act out or they might become passive. They start failing their courses. The reason they don’t get – high school sort of pushes you out. If you don’t pass enough courses, you don’t get promoted to the 10th grade; you are told to repeat the ninth grade, which is basically, like, do it over again, without anything changing, which is – and most kids actually will try to do that; they don’t drop out at that point.
So, example: We study multiple cohorts of
And if you look at those kids, you go back, this is not the first year they are waving their hands; they have been waving their hands for a bunch of years; their attendance of 80 percent the year before. They have been failing courses. These kids, in fact, have been waiving their hands since sixth grade. So I think that we really have to think about these two different segments because if you just focus on one or the other, you go in the wrong direction. And as far as interventions, I think given it’s very predictable – and I agree that we don’t have the evidence and we need to do the research to get the evidence.
But I argue, in that in the short term, there is a thing you can use to tell you, are you on the right track, but you can synthesize it from the programs you found, which is, is this intervention effectively improving attendance. Is it improving behavior? Is it decreasing course failure? Is it leading to more kids being promoted on time, because those are the mechanisms that lead to dropping out? If the intervention is changing those – and you can just look at prior years before I was in, to give you some rough indication; I’m not saying final – if that gives you some indication that you’re on the right track and that it’s not – we don’t know what to do.
MR. DYNARSKI: I don’t have any disagreement with that. I completely agree that there are two classes here, sort of the – as Bob defined it, there is a group that is reasonably capably predicted as these kids are clearly dropouts; they have signaled it. And then there is a set, which is behaving more idiosyncratically. But I think possibly school district by school district, the proportions change. That makes it slippery and certainly deserves more analysis. I’m wondering, though, whether the same program can’t actually triage to kids and develop approaches, which essentially say that you’re a long-termer, so to speak. You know, you have academic issues from way back and we need to have an academic track of recovery versus, yours is an idiosyncratic; you can do okay, but we need to mediate your relationship with the teacher. And so – (audio break, tape change) – handle a problem which is probably far more subtle than that.
MR. MURRAY: Frank Murray.
MS. WEINGARTEN: Thank you.
MR. MURRAY: I want to help you get an A. (Laughter.)
Are there schools or districts that have incredibly low dropout rates, or very high – and/or very high graduate rates, and what are their features? And part of it is to try to frame the question – (inaudible) – well, what would be a reasonable dropout rate or graduation rate, because there has got to be some boundary. And it’s coupled with – and some of these figures you look at, these don’t look like irrational decisions entirely. So you say, what would explain the Hispanic dropout rate being so much greater than – (inaudible). Well, if everybody is speaking a language I don’t get, why would I hang around? I mean, it’s really – almost nobody goes to a place where you can’t understand a language.
So if you kind of kind of look at the old outlier schools model and say there must be some places that have incredibly impressive graduation rates and very low dropout rates. So in the country, where are they? And what are the features of those places? So we know?
MR. BALFANZ (?): Well, I mean, I think you have to really do it – taking sort of affluence and poverty into account. I mean, if you just want to look nationwide, affluent suburbs, everyone graduates.
MR. MURRAY: Is that true?
MR. BALFANZ (?): Yeah. I mean, 90 percent – in the 90s. I mean, and then there is always the life-events kids that are always going to –
MR. MURRAY: So that kind of tells us that a 90-percent
graduation rate would be about as good as it’s going to get in the
MR. BALFANZ (?): Right, and especially – think about – I always argue this, is that – unless we make the legal dropout age 19 nationwide, you can never aim for better than 90 percent because kids can make the independent decision to walk away.
MR. MURRAY: So that is a reasonable ceiling.
MR. STAFFORD: That is what – yeah.
MR. BALFANZ: (Off mike.)
MR. STAFFORD: Okay, good. I wanted to get in before you left.
MS. WEINGARTEN: What is your name?
MR. STAFFORD: Oh, I’m sorry. Julian Stafford.
MS. WEINGARTEN: Thank you.
MR. STAFFORD: Julian Stafford from DCTS. I wanted to get in here before we left – that
question that was posed by the gentleman down there about where there are
models of success in terms of graduation rate.
And I would have been remised if I hadn’t mentioned
They were the first state that came
out with the partnership with colleges, where kids could transition from the 10th
grade into the college system and receive dual credit towards – you know, amassing credits and at the same time, graduating from high
school. And also, with the influx of the
immigrants from the Vietnam War,
And then with the charter school,
Another question, since I’m here, of course I want to ask – I know you have looked at – you have done a lot of studies in the urban areas in terms of programs that can be seen as innovative in terms of driving the graduation rate. I was just wondering, did you look at some of the things that are successful? What are some of the things that the private schools are doing to try to rotate those schools in the public school system? And I know of situations where students have left the public schools, have gone to private schools – students who were considered at risk, have been a tremendous success. And I was just wondering, why don’t we commission some studies to look at those types of variables to try to replicate those concerns in the public schools in terms of looking at a study?
MR. BALFANZ (?): You know, that – unfortunately, I can’t help very much about private schools, but let me just comment, almost nobody studies private schools. If a private school had an effective intervention, I don’t think I have ever heard of it. Most of what we are talking about here is mostly urban public education in fact, not even public education. So that is just another gap in the literature.
MR. STAFFORD: Just to follow up, I guess it’s sort of the
same question. I’m not really sure how
to pose it, but I’m thinking about what Russ Rumberger
said about dropping out and being a life-course process. And so I do know of a school – I mean, I
think, start early, like elementary school, there is a really good school and
it’s Linux (ph) Elementary School right near Los Angeles International Air
Port. It’s almost 100-percent Hispanic
immigrant, but it’s one of the best elementary schools in the state of
And I’m thinking more or less about the school effectiveness, school quality literature, in terms of you have these programs that you mentioned, eight programs, but to what extent, what is the context of the quality – the culture of that school that these programs are kind of dumped into. So I haven’t – I didn’t see anything in here other than family parental involvement, but to what extent does the context, the culture of the school determine the effectiveness of the program, and how does that have a net effect on student outcomes?
MR. DYNARSKI (?): Well, unfortunately, it’s hard to study the effectiveness of programs with and without that context, which is how a researcher would go about figuring out who valuable it is. But let me mention, Linux, by the way, has been well-known for a long time for being this beacon of just about everything. What – let me just speculate about what I think is happening here with this contextual factor. It’s somehow in the school’s culture, they are imparting a desire to learn, a motivation to learn.
They are – you know, we all sat through geometry or algebra as high school students, and I’m just assuming that, given the nature of the people in this room, we all thought it was okay to learn this stuff, that even if it was –
MS. WEINGARTEN: Speak for yourself. (Laughter.)
MR. DYNARSKI (?): Well, okay, yeah, but the whole notion – what I’m trying to say is that we actually bought into the promise that this all meant something somehow in some down-the-road kind of way, that I went somewhere. But you don’t have to buy this promise, by the way. You don’t – if you live in a neighborhood that is riven with violence and drug use, and where you don’t have any role models at all, how do you know that geometry actually means anything. And why should you buy into this promise? You don’t see much evidence that the promise actually gets you anywhere.
So in these – I think in these beacon schools, they have actually somehow imparted in these kids. And I think KIPP schools have this same kind of feature, by the way, that the kids actually want to learn, even though they come from settings where it is not natural that they necessarily would. So this trick, this magical bullet of being able to actually impart a wanting a learn – and by the way, all of us who have had kids through the schools – and we know the teachers who can do this too, the ones who are just – somehow they have got this – they have got it down; the kids want to learn it.
Now, you can have great curriculum, spend a lot of money on the schools, great textbooks. Actually, it wouldn’t matter at all if the kids didn’t want to learn it because you can’t force a kid to learn things; they have to do it. So I’m – I go back to this a lot in my head, because I’m trying to understand how programs can actually impart motivation. It’s not obvious how you do it in a sort of formulaic way. It’s not obvious that you can put it into a black binder and distribute it. But to the extent that you can, you should. I just don’t know the answers.
MS. WEINGARTEN: This side first, and then we’re going to go back to the back.
MS. CORTESE: Yeah, I was just wondering.
MS. WEINGARTEN: But you have to identify yourself.
MS. CORTESE: Toni Cortese, AFT.
MS. WEINGARTEN: (Off mike.)
MS. CORTESE: The Gates Foundation put a lot of money into creating small high schools in the hopes that there would be more adult-to-student contact and everything else. And I don’t know at what level of research you would put it, but I think it was generally it was – sometimes it worked, and sometimes it didn’t.
So, yeah, I just wonder, you know, at – from what we know right now, do we go around and say that the teachers are the best point of entry and that there ought to be someone that the teacher can talk to who can access certain, whether they are outside social services or academic services for a student. This seems to show that that is kind of key. And in most of our high schools, the guidance counselor is not equipped to do that, that they are either too busy – what you say, doing the college applications. But would that seem to be the best bet for someone – all of us who are involved in education to begin to concentrate on, that the person has an adult that can access them something, not just, as many high schools do, and junior high schools do, middle schools. They assign an individual.
My daughter was lucky enough to get the school secretary. I don’t know that – they are meeting once a week and imparted a lot of motivation. But would that be a good way to go. Are there any models out there that may have –
MR. BALFANZ (?): Check & Connect. Right, so in Check & Connect, it’s almost like a case manager who – that’s is probably the wrong term, that these are not necessarily social workers, but this adult has 40 kids, say. And their job is to ensure that these 40 kids are progressing in school. And various kinds of impediments will come up. Kids get sick. The case manager checks, is this a real illness, or is there a sudden loss of enthusiasm for school possibly because of breakup with a boyfriend or not doing well in a class or something like that. But it’s a diagnostic approach, trying to cut through the kind of gray area that – because, otherwise, as Bob pointed out, when kids don’t – when kids – kids drop out by not showing up anymore. There is no trumpets, there is no signing of forms or anything like that. One day they are not there.
And the problem is, if they decide to take a little bit of vacation just because things are a little bit overwhelming, you enter the vicious spiral. The vicious spiral is, then you’re a week behind, and now you are in even worse shape. So it’s harder to go back then. But if you’re two weeks behind, it’s really hard to go back. And the next think you know, you realize you dropped yourself out.
So there is this ability intervene very quickly. So-and-so has not – Johnny has not been here for three days in a row; I’m going to go figure out what is wrong with Johnny. But that is – this is not cheap. When you think about it, this person is – with a 40-kid caseload, this person is probably working full time to do problem resolution. And if you had 400 of these kids in your school, you are employing 10 people like this. You’re spending $600,000 every, if you actually imagine it with fringes. So it’s not cheap. But if you –
MS. CORTESE: (Off mike.)
MR. BALFANZ (?): Well, I think it probably is more effective, I think it’s the check-and connect model. The alternative, though, is that some proportion of those kids are going to drop out and cause the problem that brings us to the table here. So it’s not necessarily better to ignore it either. But somebody has to come up with the resources.
MR. DYNARSKI (?): I would just add one thing to it, is that – I think that is really important. But I think sometimes we get in trouble by trying to think that one lane is a solution, and I really think it’s a comprehensive approach. And my feeling is, if you look at a score form or a dropout prevention that hasn’t work, it’s a place that has bet the ranch on one leg of what has to be a three-leg school.
So we need strong instruction in the classrooms. Most kids that – many kids that drop out are academically behind, and you can’t just sort of gloss that over, and you have to have strong, effective instruction, and really effective extra help. I mean, this is one thing that drives me crazy. We spend so much money on tutoring and extra help, and it’s so uncoordinated in practice.
I mean, what typically happens is it – the tutor will hear, Johnny is struggling in math. That is what they will know. And they will talk to Johnny and they’ll give a diagnostic test, and they might say, hmm, you’re struggling with decimals. I’m going to teach you decimals. But, in fact, what he is studying the classroom is proportionality, and that is what his test is in. And if he doesn’t get help in proportionality or something else, he is going to fail that test, and that is what is going to put him behind.
So that to me is another huge area of why I think sometimes we don’t have effective studies, is because we have the basic idea right, but we don’t put it together right, which is, if the extra help doesn’t help you pass your course, just generally making your skills better is not going to solve the problem. I mean, it’s good to make your skills better, but it’s not going to solve this problem.
And then we do need this social support and this engagement. I mean, the best instruction in the world won’t work if the kid doesn’t come and try and behave. And you have to engage kids, especially when there are so many counter-pulls to their lives, especially in high-poverty environments. I mean, 12-year-old girls are asked to stay home to take care of siblings. How do you deal with that counter-pressure? You know, how do you deal with that they have to take a longer distance to get to school, and it takes them through crime areas, and they have to avoid that. I mean, that puts a lot of stress on the kid. And if you don’t acknowledge that, it festers.
And then finally, we’re asking teachers to do so much more, right. Teach a great lesson with kids at all ability ranges, and call them when they are not in school. So if we don’t build the support around the teacher, it becomes dependent upon a few heroic folks that do it for two years and burn out. And so you have got to build the supports around the teachers to – so another thing I have seen that there is not the research on it yet, but it’s a variation of Check & Connect, and it’s better than having your one lone person take 10 kids, whatever there problem is, try to solve it, even if it’s totally out of their skill set, is to have teams of teachers do that.
And if it is a team of four adults sharing 20, 30, 50 kids, they can go to your strengths. I’m good with parents, so I’ll deal with parents. I’m good with tutoring so I’ll arrange tutoring. But then those teachers have to have time to meet together, right, to do that, and it’s got to be built into their job; it can’t be, like, stay after school and do it.
So I think we have to try some comprehensive models, based on what we think is important, and then test them to prove that they work.
MS. ROLLIN: Miriam Rollin, Fight Crime: Invest in Kids. We are very interested because our police chiefs and sheriffs and DAs end up with a lot of these kids that drop out, so we are trying to reduce the number of clients that our members get.
One evidence-based approach that wasn’t mentioned in terms of dropout prevention, but I want to throw there is pre-K, Perry (ph), CTC (ph), et cetera – evidence showing reductions in dropout and increased graduation rates. But I wanted to ask, this kind of says which ones of these had effects and which ones didn’t, and I was wondering if you could comment at all on any differentials and effect sizes. So this one was really good, and this one had a little effect. If you could comment on that at all.
And then also – actually, you mentioned this, Bob. What about simply increasing the age of mandatory attendance. I know that has been tried in some places, but I don’t know what the impact of that is.
MR. BALFANZ (?): Yeah, I mean, you can – there is in fact –
each of these interventions has a very long description about them on the web
page, but let me just do a thumbnail sketch.
So, yes, they worked, but – and
they were tested and found to work, but not in any – not in the kind of
dimension that Check & Connect and
Up there in that range of distribution, that is actually a really large fact. So 15 percent of the control group just fell away, or 17 percent, and almost none of the treatment group did, so that was quite dramatic.
I’m sorry, did you have another question? There was another question.
MS. ROLLIN: Well, talking about just increasing the mandatory school attendance to age 18. I mean, I know some places have experimented with that.
MR. DYNARSKI (?): Yeah, I mean –
MS. WEINGARTEN: State by state.
MR. DYNARSKI (?): In one place I have seen this, and it’s sort of inconclusive to me, is that Pennsylvania and Maryland are off by a year, and one, you can drop out at 16; on you can drop out at 17. What that basically does, is there is more repeaters in one than the other, but the dropout rates are about the same, which means I think it does – it does keep kids in school a little bit longer. I’m not sure if it’s leads them to graduating more. So, I mean, it is sort of – for some kids, it’s enough of deterrent that you legally can’t leave, that they’ll still come a little bit longer, which gives you more time, if you’re going to do something about; if you’re not going to do something about it, I don’t think the law itself will have that huge effect.
MS. ROLLIN: (Off mike.)
MR. DYNARSKI (?): Yeah, there is other – some states have that, I guess; I’m just – I’m not aware of the study.
MS. : (Off mike.) And I kind of want to complicate the picture a little bit more, and that is to say, we’re talking about getting kids to the end of graduation, to actually graduate. And believe me, that is important. But I would challenge us to consider that what they are graduating with and the preparation they have is not necessarily sufficient, and that if we want to be comprehensive in how we think about this, there is really kind of a dual agenda. It’s raising the academic standards so that it is connected and anchored in what students can expect when they graduate in the real world, in college – technical training – two-year program, four-year program in the workplace.
I think increasingly the message is that there is this growing recognition that a high school diploma is not enough. So we’re trying to get kids to get the diploma, but I want to make sure that the diploma we get them with has enough value, and that if we’re going to take on this challenge, maybe we should take it on comprehensively. And I didn’t know if you wanted to comment on that.
MR. DYNARSKI (?): Well, I’ll – sorry go ahead.
MS. WEINGARTEN: Let me just – we’re seeing a lot of crosscurrent sail on that, in that with AYP and with testing, I’m actually seeing that curriculum is getting dumbed down in a lot of places because there is a need to pass, like, we’ll take the Regents in New York, just in terms of my very – not my analysis based upon a researcher, but on my analysis based upon a former social studies teacher. The government and American history Regents these days are a far cry from what they were even three, four, five years ago.
And so part of the dilemma, when you raise both questions at the same time, is that what are the standards looking like as they are described by tests, and whether the testing is actually outdistancing everything else that kids need to know and be able to do in terms of high school.
So I just – I want to just add a cautionary note here in terms of raising that issue at the same exact time as we’re raising the issue about graduations and dropouts.
MR. DYNARSKI (?): Let me comment – just to add a fact here. NCES put out a report in September where they compared the characteristics of 10th graders in 1980, 1992, and 2002. So basically you have a 22-year span where the 10th grader can change. And there is a fascinating chapter where they point out that the proportion of kids who now say they are in a college preparatory curriculum doubles in 22 years. And in the lowest socioeconomic stratum, it triples from about 20 to 60.
So it goes to the – it’s indirectly around your issue, but it’s basically, the kids – presuming – I’m presuming that a college pre-prep curriculum was better than what they would otherwise have had in the general track, so they moved up in some kind of academic sense. Now, they might have moved. Randy’s other point is, but they might have used the college prep curriculum down.
MS. : Well, I’m the first to say course title is not the same thing as content.
MR. DYNARSKI (?): Right, right. But on the other hand – and what corroborates this is that the proportion of kids who say that their parents expect them to go to college immediately after high school also went up commensurately, which means that they also sense that there is this greater pressure that you mentioned for going on to college. We would be – we should be concerned of course if their academic preparation is actually going in the opposite direction. But test scores did not change over the 22 years, the ones that are given inside high school and beyond and now. In fact, they were higher in math, and very, very slightly lower in reading, but the kids weren’t do a lot worse surprisingly.
MR. : Yeah, we took it over here. We stole it.
MR. : I have two quick comments. One is that the states themselves define what
is meant by a high school diploma, and there are very significant differences
across the states. So, for example,
like, we have done some examples on that, and my favorite example is that of
But when you look at the net
scores, you find that in almost all of the categories, like fourth-grade
reading, fourth-grade mathematics, eighth-grade reading, eighth-grade
mathematics, blacks in
And the other thing is that I would like something that was mentioned, particularly by Bob again and again, that – I had a student of mine who was working on the – (inaudible) – ’97, and they had, like, a big national sample of people who were basically 12 to 16 year olds in 1996, and they followed them, interviewed them every year. And when you look at their high school completion, you can control many variables, like their sex, their – like, sex, racial – like their race, their parental income, parents’ education and everything. But the one thing that comes up to be most significant are middle school grades, and simultaneously, like, whether they had – (inaudible) – a grading school.
So these particular things swamp everything else like middle school grade and whether they have – (inaudible) – a grade all year.
MS. SAMUELS: Christina Samuels, Education Week. You were – I was looking, obviously, as you were saying, there is sort of a small number of studies over these 20-year window that you were looking at. I was wondering if there are – or if you know if there are more studies that are sort of underway or in the works, or if they are – you know, you were looking – over 20 years. I wonder if there was, like, more towards the latter end of that, which means that there might be more research that is being done in this, or is it still of the rigor and study that you were looking at still just sort of spotty over that 20-year period.
MR. DYNARSKI (?): There have – I think it’s actually increasing recently. Let me just sketch a picture. Starting about 1988 with the Humphrey-Hawkins Bill, there was federal attention on dropout prevention inside the legislation. So they have created a program; they began studying it. There were two waves of something called the school demonstration assistance program. I was part of the study of the second one, which was – used experiments and the like. So I was very familiar with it. It ended about ’97.
Starting around – going back a couple of years prior to that – research often is a lagging indicator of what is going on – even two years prior to that, the federal government had essentially nearly ended its support for dropout prevention demonstrations. It had basically seeded the field to the states in the local governments.
There are a lot of programs out there. I think there are a lot of evaluations of these programs out there. There is very little investment in them, however, in the sense of effectiveness research. They are mostly reasonably small-scale studies that actually just don’t meet the scale of rigor that the federal government often requires.
And then later in the ’90s, partly through I think the Gates money flowing out there, but partly because Johns Hopkins and Mend (ph), DRC, and other organizations kept the flame burning, to speak. The Talent Development Model was launched. First Things First research began at that time.
So I think there are some very promising signs recently about – basically, people have kept us going. I don’t know – I mean, frankly, I’m not sure why. The federal government’s dropout prevention program was $5 million last year. Five million dollars? Isn’t that a rounding error? (Laughter.) So it’s just – why would somebody – at least at the level at which you can actually underwrite large-scale research, $5 million is never going to do it.
MS. WEINGARTEN: There is a bill, though, in the Senate, isn’t there, that Kennedy is pursuing, and what would that – that may go to the question that the – (inaudible) – asked as well. If that bill got passed, would it help the situation?
MR. BALFANZ (?): Yes. (Laughter.)
MR. DYNARSKI (?): Well, partly it calls for an explicit – it would help the research agenda certainly because it calls for an explicit set-aside for research and capacity building.
MR. BALFANZ (?): Ninety-six million.
MR. DYNARSKI (?): Yeah, now – oh, yeah, that is a lot. Got my attention.
MR. BALFANZ (?): Four percent – 2.5 billion.
MS. WEINGARTEN: That is not a rounding error. (Laughter.)
MR. DYNARSKI (?): That is not – no, that is a lot. So basically it’s been pretty undernourished area, as I mentioned.
MS. WEINGARTEN: Can I – it’s close to 2:30, and I’m supposed to keep everything on time. So is there one more burning question that hasn’t been asked? (Laughter.)
MR. BALFANZ (?): Flip a coin.
MS. WEINGARTEN: So then, since there is two more burning questions, in the tradition of Albert Shanker of being democratic, we’ll do both – (chuckles). Howard, and then Frank, right?
MR. : Gerald.
MS. WEINGARTEN: Gerald. Okay, and then if the two of you gentlemen would like to also quickly give your parting wise words to us after Howard and Gerald, we would love that as well.
MR. NELSON: Howard Nelson, the AFT. I took a good look at the second graph, and it looks to me like the high-school completion rate increased for black and Hispanic students by about 15 points, which, for black students, cut the gap in half, which I think is pretty significant. And that leads into my second related question I have, and that is that GED has been mentioned disparagingly here, but the data show that the percentage of high school graduates that are GEDs has risen from 2 to 22 percent during this time period.
So why aren’t we talking about GEDs more plus or minus, because there is a lot of interactions here, good or bad. Like, it’s good that people are getting GEDs, but maybe because kids can get GEDs now, they are forced out of schools, and maybe schools themselves are encouraging GEDs. And charter schools are encouraging GEDs because at high school age, a lot of programs are aimed at dropouts to get them a diploma with GED. So it’s an alternative route to high school for high school-age kids. And you can go on and on and on, but there is a whole different subject here that integrates into this discussion.
MR. BALFANZ (?): Two things. One is Tom Caine did a nice analysis, where he basically demonstrates that the increase in the black completion rate is very much attributable to increases in the black – in parents of black students having higher education themselves. But I agree with you that that definitely went up and it was a very good thing, but then it tempered off. Basically, after 1990, it kind of – so it went up like that and then stopped, which was – we would like to see it keep going.
With respect to the GED, I would be more than happy to have a discussion about it. I don’t think it necessarily should be the unfortunate stepsister sort of thing to the diploma because for kids who are 17 and 18 and they really don’t have any credits yet in their high school courses, I don’t see a realistic alternative. You can’t really expect a kid at that age to sit in high school for five more year; they won’t do it. So having the GED as a realistic alternative makes sense to me; however, I think it needs to not be terminal, and that is a key thing. And we wrote about this in some paper. (Laughter.)
But the idea is we – what we essentially said was GED should be integrating with post-secondary systems so that a kid can receive a GED and keep going because, frankly, if you have an associates degree or a training certificate or something like that, your employer is much less likely to care about what your high school completion status was. But if it doesn’t integrate, then you become a GED completer, and that is a stigma I think in the labor force.
Let me also mention that I – in the slides I pointed out, that GED programs were not yet explored by the what-works clearing house, but we already know that there are programs out there which can dramatically increase GED completion rates. They are in the welfare system; they are in a job-training system; they are in the education system. Virtually all of these different kinds of programs have been found to increase GED completion rates. Job Corps, for example, increases it dramatically.
So we can in fact get kids a GED. Do we want to? You need a forum for that. (Laughter.)
MS. WEINGARTEN: I hear a call for another forum.
MR. BALFANZ (?): Another forum.
MR. SHROUFE: Gerry Shroufe with AERA. And I had – I wanted to support the idea of a lengthier required high school education. Researchers at NIH are convinced that the brain development is such that judgment is the part of the brain that develops most recently, last. And they are even advocating raising the age of driving to get a drivers’ license because they fear that the judgment that we assume because people look mature simply isn’t there, and it is particularly true of males. (Laughter.)
The question I had, though, was whether or not, given your obvious commitment to the issue, you think we can actually address this through the education system, or ought we be looking at the welfare employment system to resolve these problems.
MR. BALFANZ: I’m going to grab that because I have to give a quick answer, and then unfortunately I have to go because I have to get back for a guitar recital for my son.
But, yes and no is a frustrating thing. Obviously we need to involve the other welfare and social service agencies. But the schools are vitally important. And there is too many schools that are just under-resourced and overwhelmed that are essentially pipelines to dropping out, not for lack of trying or lack of caring but because we have concentrated the neediest students in a subset of schools that are just overmatched. You know, if we don’t get that right, it’s much more expensive to use social services and health reform, which we vitally need. So, yes, we need to do it, but I really think it’s – you have to really keep the laser-like focus on the key school reform points where we can really make a difference.
And I guess my last thought, parting thought is that I think it’s really important – getting back to grad-rate measurement – than the reauthorization of No Child Left Behind. We finally get it right, right. We need to have a common goal across al states. It should be that 90 percent at least. We need to have some sort of common growth rate, common expectations to get there, and then make sure we get through things like the Graduation Promise Act, and other legislation that is out there, the resources to these schools, and sort of the – my personal opinion now is the tradeoff is, you get the resources you need but you’re held accountable for implementing what we know works, and what we have at least promising evidence that works, even if we don’t have final evidence.
And on that, I think you for having us here, and hope the dialogue continues.
MS. WEINGARTEN: Mark, do you have any parting comments?
MR. DYNARSKI: I have none. Thank you. That was very well done. (Laughter.)
MS. WEINGARTEN: Thank you, John. Thank you, Laura. Thank you, all of you, before being here to day at this seminar.
(END)